In this article, we’ll discuss the possibilities of real-time anomaly detection in smart cities and how it can enable us to build a safer future.
Smart cities are well connected and data-driven urban areas, that deliver services and infrastructure to elevate the living standards of residents and facilitates a seamless environment for running businesses, city operations, and management. With the onset of climate change, economic instabilities due to the COVID-19 pandemic, and the shooting population growth, there are greater demands from the public for smart living conditions that are safe, resilient, and highly responsive to anomalies or disruption. And to achieve that, the city infrastructure has to be digital, interconnected, intelligent, and futuristic. Human aspirations have to be met in every aspect – whether it’s smart traffic management, disaster management readiness, or crime prevention.
The best strategy for crime prevention for any city’s police force is to intervene as quickly as possible, and with as much information as they can gather. For example, the city of Detroit began Project Green Light in 2016 with exactly these objectives, the Detroit Police Department accessed CCTV cameras around partner locations (mostly gas stations where the crime rate is high) to transmit live footage to PD headquarters, where it can be monitored by a 40-member police unit. This footage helps officers catch criminals, the project combines real-time crime-fighting and community policing to improve neighborhood safety and promote the growth of local businesses. It was announced in 2017 after the implementation of Project Green Light, violent crime has gone down by up to 50 percent. Even though the police department was able to reduce the crime rate by 50%, they had to allocate a 40-member team for that purpose. This is, most probably not a feasible option for the majority of police departments that are understaffed in different cities around the world. Moreover, the average attention span of humans is about 8 seconds, which increases the probability of missing out to detect anomalous activities with manual monitoring of CCTV footages. Can this process be automated?
Technology’s advancements and its applications transform a city to a smart one. Advanced security and safety of the residents is a top priority for smart cities. At present, most cities in the world are heavily under surveillance using CCTV cameras that are mostly monitored by CCTV operators. Humans are simply not efficient enough for the growing demands of surveillance security monitoring to identify rare events across multiple video streams. In a 2014 Harvard Medical School study in visual screening tasks similar to that done by CCTV operators, participants made frequent errors in real-time decision making. These errors were almost always false anomalies, most people failed to detect the target as it passed their field of sight. Psychological phenomena reduce people’s efficiency to detect anomalies especially at jobs that require constant monitoring of objects for a rare occurrence. In high-security institutions such as banks, jails and courthouses it could prove devastating to oversee these grave errors in human judgment.
Contact us for a demo
AI-powered Video analytics systems can transform normal surveillance cameras into a crucial security infrastructure. Anomaly detection is helpful to detect anomalies happening in a particular location – traffic intersections, office entry gates, parking lots, warehouses or even a neighborhood park. For example, in a parking lot anomaly detection enabled CCTV cameras will be able to identify theft, automated number plates tracking and more. It can identify fire or accidents in a warehouse, unusual movements and human presence in restricted areas of office space – with minimum human intervention.
Emotyx is an AI-powered video analytics software with a range of applications in anomaly detection. The product can be installed in CCTV cameras to identify theft, road accidents, fights, shootings, explosions, and such anomalous incidents. It can be integrated into existing surveillance systems and alert authorities to make swift decisions in real-time. When a deviation from the usual activities that happen in a place, the system will alert the concerned official, to take automated actions such as alerting firefighters in case of a fire detected in a building.
Applications of anomaly detection include Monitoring mechanical processes executions in factory premises, Identifying unusual activities that need to be alerted to officials, automated experiments, rules violation detection, etc. The video below shows the real-time anomaly detection applications of Emotyx:
With the addition of Anomaly detection in the Project green light of Detroit, the outcomes could have been quadrupled. Few other examples of smart anomaly detection use cases are the smart traffic sensors in New York and crime prediction software in Santa Cruz, California. Santa Cruz Police Department implemented predictive policing software from an algorithm-generated from about eight years of data on car and home burglaries. A city interwoven with physical devices and intelligent systems will generate data that are crucial to its growth and security. These cities demand efficient systems to help businesses and individuals to generate productive outcomes from data. One of the fundamental aspects of any ‘Smart City’ is a ‘Safe City’ – AI video analytics can deliver exactly that by transforming an average CCTV camera to a critical high-level security feature.